R Notebook

Loading the Packages

library(pacman)
#p_install_gh("favstats/favstats")
p_load(tidyverse, haven,sjmisc,
                     forcats, weights, car,
                     countrycode, lavaan,
                     semTools, lavaan.survey,
                     survey, reshape2, favstats,
                     magrittr, skimr, xray)
range01 <- function(x){(x-min(x, na.rm = T))/(max(x, na.rm = T)-min(x, na.rm = T))}

Loading the URLs

afro5_url <- url("https://github.com/favstats/database_delib/raw//master/afro5.RData")
afro6_url <- url("https://github.com/favstats/database_delib/raw//master/afro6.RData")
latino2013_url <- url("https://github.com/favstats/database_delib/raw//master/latino2013.RData")
latino2015_url <- url("https://github.com/favstats/database_delib/raw//master/latino2015.RData")
wvs_raw_url <- url("https://github.com/favstats/database_delib/raw//master/wvs_raw.RData")
americas_url <- url("https://github.com/favstats/database_delib/raw//master/americas.RData")
bolivia_url <- url("https://github.com/favstats/database_delib/raw//master/bolivia.RData")
canada_url <- url("https://github.com/favstats/database_delib/raw//master/canada.RData")
asian_raw_url <- url("https://github.com/favstats/database_delib/raw//master/asian_raw.RData")
myanmar_raw_url <- url("https://github.com/favstats/database_delib/raw//master/myanmar_raw.RData")
mongolia_raw_url <- url("https://github.com/favstats/database_delib/raw//master/mongolia_raw.RData")
philip_raw_url <- url("https://github.com/favstats/database_delib/raw//master/philip_raw.RData")
taiwan_raw_url <- url("https://github.com/favstats/database_delib/raw//master/taiwan_raw.RData")
thai_raw_url <- url("https://github.com/favstats/database_delib/raw//master/thai_raw.RData")
malay_raw_url <- url("https://github.com/favstats/database_delib/raw//master/malay_raw.RData")
singapore_raw_url <- url("https://github.com/favstats/database_delib/raw//master/singapore_raw.RData")
sk_raw_url <- url("https://github.com/favstats/database_delib/raw//master/sk_raw.RData")
cambodia_raw_url <- url("https://github.com/favstats/database_delib/raw//master/cambodia_raw.RData")
ESS_raw_url <- url("https://github.com/favstats/database_delib/blob/master/ESS_raw.Rdata?raw=true")
vdems_start_url <- url("https://github.com/favstats/database_delib/raw//master/vdems_start.Rdata")
qog_url <- url("https://github.com/favstats/database_delib/raw//master/qog.Rdata")

Afrobarometer Data

Afro 5

load(afro5_url)
delete_na_afro <- function(x) {
  x <- Recode(x, "9 = NA;
            98 = NA;
            99 = NA;
            -1 = NA")
  return(x)
} # Funktion um die NAs im Afro Datensatz zu bestimmen
# alle Variablen
afro_5 <- afro5 %>%
  rename(
    educ = Q97, # Variablen umbenennen
    income = Q3B,
    sex = Q101,
    work = Q96,
    trust_gov = Q59A,
    trust_parliament = Q59B,
    trust_police = Q59H,
    trust_courts = Q59J,
    # demtoday = Q46A
  ) %>%
  mutate_at(
    vars(
      income, trust_gov, trust_parliament,
      trust_police, trust_courts, work
    ),
    delete_na_afro
  ) %>% # NAs deleten
  mutate(
    sex = sex - 1, # sex (0/1) codieren
    work = Recode(
      work, # work (0/1) codieren
      "1 = 0;
       2 = 1;
       3 = 1"
    ),
    age = Recode(
      Q1, # missing values l?schen
      "-1 = NA;
      998 = NA;
      999 = NA"
    ),
    # demtoday = Recode(
    #   demtoday,
    #   "-1 = NA;
    #    98 = NA;
    #    99 = NA"
    # ),
    educ = Recode(
      educ,
      "-1 = NA;
       98 = NA;
       99 = NA;
      999 = NA"
    ),
    cntry = to_label(COUNTRY_ALPHA),
    year = as.numeric(format(DATEINTR, "%Y"))
  ) %>%
  select(
    cntry, year, age, sex, income, educ, work, 
    trust_gov, trust_parliament, trust_police, trust_courts
  )
afro_5

save(afro_5, file = "codebooks/singles/afro_5.Rdata")

Afro 6

load(afro6_url)
afro_6 <- afro6 %>%
  rename(
    educ = Q97,
    income = Q4B,
    sex = Q101,
    work = Q95,
    trust_gov = Q52A,
    trust_parliament = Q52B,
    trust_police = Q52H,
    trust_courts = Q52J,
    # demtoday = Q40
  ) %>%
  mutate_at(
    vars(
      income, trust_gov,
      trust_parliament, trust_police,
      trust_courts, work
    ),
    delete_na_afro
  ) %>% # NAs deleten
  mutate(
    sex = sex - 1, # sex (0/1) codieren
    work = Recode(
      work, # work (0/1) codieren
      "1 = 0;
       2 = 1;
       3 = 1"
    ),
    age = Recode(
      Q1, # missing values l?schen
      "-1 = NA;
      998 = NA;
      999 = NA"
    ),
    educ = Recode(
      educ,
      "-1 = NA;
       98 = NA;
       99 = NA;
      999 = NA"
    ),
    # demtoday = Recode(
    #   demtoday,
    #   "8 = NA"
    # ),
    cntry = to_label(COUNTRY_R5List),
    year = as.numeric(format(DATEINTR, "%Y"))
  ) %>%
  #  mutate_at(vars(income, educ, age,
  #                 trust_gov, trust_parliament,
  #                 trust_police, trust_courts),
  #            as.character) %>%
  #  mutate_at(vars(income, educ, age,
  #                 trust_gov, trust_parliament,
  #                 trust_police, trust_courts),
  #            as.numeric) %>%
  select(
    cntry, year, age, sex, income,
    educ, work, #demtoday,
    trust_gov, trust_parliament,
    trust_police, trust_courts
  ) 
afro_6

save(afro_6, file = "codebooks/singles/afro_6.Rdata")

Merging the Data

afro_real <- afro_5 %>% 
  mutate_at(
    vars(
      income, educ,
      trust_gov, trust_parliament,
      trust_police, trust_courts
    ),
    range01
  )
afro_real2 <- afro_6 %>% 
  mutate_at(
    vars(
      income, educ,
      trust_gov, trust_parliament,
      trust_police, trust_courts
    ),
    range01
  )
# length(unique(afro_real$cntry))
# length(unique(afro_real2$cntry))
# table(afro_real2$year)
afro <- rbind(afro_real, afro_real2) %>% as.tbl()
afro

Latino Barometro

Latino 2013

load(latino2013_url)
latino_2013 <- latino2013 %>%
  rename(
    educ = REEDUC_1,
    income = S6,
    # demtoday = P50TGB.A,
    age = S11
  ) %>%
  mutate(
    sex = S10 - 1, # sex variable erstellen
    cntry = to_label(IDENPA), # cntry variable erstellen
    trust_gov = 5 - P26TGB.B,
    trust_parliament = 5 - P26TGB.C,
    trust_police = 5 - P28TGB.B,
    trust_courts = 5 - P26TGB.E,
    year = 2013,
    income = 5 - income,
    work = Recode(
      S19.A,
      "2 = 1;
       3 = 1;
       4 = 0;
       5 = 0;
       6 = 0;
       7 = 0"
    )
  ) %>%
  select(
    cntry, year, age, sex, income, educ, work, 
    trust_gov, trust_parliament, trust_police, trust_courts
  )
latino_2013

save(latino_2013, file = "codebooks/singles/latino_2013.Rdata")

Latino 2015

load(latino2015_url)
latino_2015 <- latino2015 %>%
  mutate(sex = S12 - 1) %>% # sex variable erstellen
  rename(
    educ = REEDUC_1,
    income = S4,
    # demtoday = P17STGBS,
    age = S13
  ) %>%
  mutate(
    trust_gov = 5 - P16ST.G,
    trust_parliament = 5 - P16ST.F,
    trust_police = 5 - P16TGB.B,
    trust_courts = 5 - P16ST.H,
    work = Recode(
      S21.A,
      "2 = 1;
       3 = 1;
       4 = 0;
       5 = 0;
       6 = 0;
       7 = 0"
    ),
    cntry = to_label(IDENPA),
    income = 5 - income,
    year = 2015
  ) %>%
  select(
    cntry, year, age, sex, income, educ, work, 
    trust_gov, trust_parliament, trust_police, trust_courts
  )
latino_2015

save(latino_2015, file = "codebooks/singles/latino_2015.Rdata")

Merging the Data

latino_real <- latino_2013 %>% 
  mutate_at(
    vars(
      income, educ,
      trust_gov, trust_parliament,
      trust_police, trust_courts
    ),
    range01
  )
latino_real2 <- latino_2015 %>% 
  mutate_at(
    vars(
      income, educ,
      trust_gov, trust_parliament,
      trust_police, trust_courts
    ),
    range01
  )
latino <- rbind(latino_real, latino_real2) %>% as.tbl()
latino

World Value Survey

load(wvs_raw_url)
wvs <- wvs_raw %>%
  rename(
    educ = V248,
    income = V239,
    # demtoday = V141,
    age = V242,
    year = V262
  ) %>%
  mutate(
    sex = V240 - 1, # sex variable erstellen
    trust_gov = 5 - V115,
    trust_parliament = 5 - V117,
    trust_police = 5 - V113,
    trust_courts = 5 - V114,
    work = Recode(
      V229,
      "2 = 1;
       3 = 1;
       4 = 0;
       5 = 0;
       6 = 0;
       7 = 0;
       8 = 0"
    ),
    cntry = to_label(V2)
  ) %>%
  select(
    cntry, year, age, sex, income, educ, work, 
    trust_gov, trust_parliament,
    trust_police, trust_courts
  )
wvs

save(wvs, file = "codebooks/singles/wvs.Rdata")

Americas Barometer

Americas Main

load(americas_url)
americas_ <- americas %>%
  mutate(cntry = to_label(pais)) %>%
  select(
    cntry, ocup4a, ed, q10new, q1, q2,
    b10a, b13, b18, b21a, n3, year
  ) %>%
  rename(
    work = ocup4a,
    educ = ed,
    income = q10new,
    sex = q1,
    age = q2,
    trust_courts = b10a,
    trust_parliament = b13,
    trust_police = b18,
    trust_gov = b21a,
    # demtoday = n3
  ) %>%
  mutate(
    sex = sex - 1, # sex variable erstellen
    work = Recode(
      work,
      "2 = 1;
       3 = 1;
       4 = 0;
       5 = 0;
       6 = 0;
       7 = 0"
    )
  ) %>%
  select(
    cntry, year, age, sex, income, educ, work,  
    trust_gov, trust_parliament, trust_police, trust_courts
  )
americas_

save(americas_, file = "codebooks/singles/americas_.Rdata")

Bolivia

load(bolivia_url)
bolivia_ <- bolivia %>%
  mutate(cntry = to_label(pais)) %>%
  select(
    cntry, ocup4a, ed, q10new, q1, q2,
    b10a, b13, b18, b21a, n3, year
  ) %>%
  rename(
    work = ocup4a,
    educ = ed,
    income = q10new,
    sex = q1,
    age = q2,
    trust_courts = b10a,
    trust_parliament = b13,
    trust_police = b18,
    trust_gov = b21a, # viele missing values
    # demtoday = n3
  ) %>%
  mutate(
    sex = sex - 1, # sex variable erstellen
    work = Recode(
      work,
      "2 = 1;
       3 = 1;
       4 = 0;
       5 = 0;
       6 = 0;
       7 = 0"
    )
  ) %>%
  select(
    cntry, year, age, sex, income, educ, work, 
    trust_gov, trust_parliament, trust_police, trust_courts
  )
bolivia_

save(bolivia_, file = "codebooks/singles/bolivia_.Rdata")

Canada

load(canada_url)
canada_ <- canada %>%
  mutate(cntry = to_label(pais)) %>%
  select(
    cntry, exc13, education, q10, q1, q2,
    b10a, b13, b18, b21a, n3, year
  ) %>%
  rename(
    work = exc13,
    educ = education,
    income = q10,
    sex = q1,
    age = q2,
    trust_courts = b10a,
    trust_parliament = b13,
    trust_police = b18,
    trust_gov = b21a,
    # demtoday = n3
  ) %>%
  mutate(
    sex = sex - 1, # sex variable erstellen
    work = work - 1,
    income = ifelse(income == 88, NA, income)
  ) %>%
  select(
    cntry, year, age, sex, income, educ, work,  
    trust_gov, trust_parliament, trust_police, trust_courts
  )
canada_

save(canada_, file = "codebooks/singles/canada_.Rdata")

Merging the Data

americas <- americas_ %>% 
  mutate_at(
    vars(
      income, educ,
      trust_gov, trust_parliament,
      trust_police, trust_courts
    ),
    range01
  )
bolivia <- bolivia_ %>% 
  mutate_at(
    vars(
      income, educ,
      trust_gov, trust_parliament,
      trust_police, trust_courts
    ),
    range01
  )
canada <- canada_ %>% 
  mutate_at(
    vars(
      income, educ,
      trust_gov, trust_parliament,
      trust_police, trust_courts
    ),
    range01
  )
americas <- rbind(americas, bolivia, canada) %>% 
  mutate(cntry = as.character(cntry)) %>% 
  mutate(cntry = ifelse(cntry == "Haití", "Haiti", cntry))
americas

# 
# table(americas$cntry)

Asian Barometer

load(asian_raw_url)
delete_na_asian <- function(x) {
  x <- Recode(
    x,
    "-2 = NA;
      7 = NA;
      8 = NA;
      9 = NA;
     97 = NA;
     98 = NA;
     99 = NA;
     -1 = NA"
  )
  return(x)
} # Funktion um die NAs im asian Datensatz zu bestimmen
asian_3 <- asian_raw %>%
  rename(
    educ = se5,
    trust_gov = q9,
    trust_parliament = q11,
    trust_police = q14,
    trust_courts = q8,
    sex = se2,
    income = se13a,
    work = se9
  ) %>%
  mutate_at(
    vars(
      income, trust_gov, trust_parliament,
      trust_police, trust_courts, work, sex
    ),
    delete_na_asian
  ) %>% # NAs deleten
  mutate(
    trust_gov = 5 - trust_gov,
    trust_parliament = 5 - trust_parliament,
    trust_police = 5 - trust_police,
    trust_courts = 5 - trust_courts,
    sex = sex - 1, # sex (0/1) codieren
    income = Recode(
      income,
      "0 = NA"
    ),
    income = 5 - income,
    work = Recode(
      work, # work (0/1) codieren
      "2 = 0"
    ),
    age = Recode(
      se3a, # missing values l?schen
      "-1 = NA"
    ),
    # demtoday = Recode(
    #   q91,
    #   "-1 = NA;
    #    97 = NA;
    #    98 = NA;
    #    99 = NA"
    # ),
    educ = Recode(
      educ,
      "-1 = NA;
       98 = NA;
       99 = NA"
    ),
    cntry = to_label(country),
    year = as.numeric(format(ir9, "%Y"))
  ) %>%
  select(
    cntry, year, age, sex, income, educ, work, 
    trust_gov, trust_parliament, trust_police, trust_courts
  ) 
asian_3

save(asian_3, file = "codebooks/singles/asian_3.Rdata")

Wave 4

load(myanmar_raw_url)
needed <- function(data) {
  ss <- data %>%
    rename(
      educ = se5,
      trust_gov = q9,
      trust_parliament = q11,
      trust_police = q14,
      trust_courts = q8,
      sex = se2,
      income = se13a,
      work = se9
    ) %>%
    mutate_at(
      vars(
        trust_gov, trust_parliament,
        trust_police, trust_courts, work, sex
      ),
      delete_na_asian
    ) %>% # NAs deleten
    mutate(
      trust_gov = 5 - trust_gov,
      trust_parliament = 5 - trust_parliament,
      trust_police = 5 - trust_police,
      trust_courts = 5 - trust_courts,
      sex = sex - 1, # sex (0/1) codieren
      income = Recode(
        income,
        "0 = NA;
        97 = NA;
        98 = NA;
        99 = NA;
        -1 = NA"
      ),
      work = Recode(
        work, # work (0/1) codieren
        "2 = 0"
      ),
      age = Recode(
        se3_2, # missing values l?schen
        "-1 = NA"
      ),
      # demtoday = Recode(
      #   q94,
      #   "-1 = NA;
      #    97 = NA;
      #    98 = NA;
      #    99 = NA"
      # ),
      educ = Recode(
        educ,
        "-1 = NA;
         98 = NA;
         99 = NA"
      ),
      cntry = to_label(country),
      year = year
    ) %>%
    select(
      cntry, year, age, sex, income, educ, work, 
      trust_gov, trust_parliament, trust_police, trust_courts
    )
  
  return(ss)
}
myanmar <- needed(myanmar_raw)
load(mongolia_raw_url)
load(philip_raw_url)
load(taiwan_raw_url)
load(thai_raw_url)
load(malay_raw_url)
load(singapore_raw_url)
load(sk_raw_url)
load(cambodia_raw_url)
mongolia  <- needed(mongolia_raw)
philip    <- needed(philip_raw)
thai      <- needed(thai_raw)
malay     <- needed(malay_raw)
singapore <- needed(singapore_raw) # Singapore hat extrem viele Missing values
sk        <- needed(sk_raw)
cambodia  <- needed(cambodia_raw)
save(myanmar, file = "codebooks/singles/myanmar.Rdata")
save(mongolia, file = "codebooks/singles/mongolia.Rdata")
save(philip, file = "codebooks/singles/philip.Rdata")
save(thai, file = "codebooks/singles/thai.Rdata")
save(malay, file = "codebooks/singles/malay.Rdata")
save(singapore, file = "codebooks/singles/singapore.Rdata")
save(sk, file = "codebooks/singles/sk.Rdata")
save(cambodia, file = "codebooks/singles/cambodia.Rdata")

Sonderfall Taiwan

# Taiwan Problem: Income only has 3 categories
taiwan <- taiwan_raw %>%
    rename(
      educ = se5,
      trust_gov = q9,
      trust_parliament = q11,
      trust_police = q14,
      trust_courts = q8,
      sex = se2,
      income = se14a,
      work = se9
    ) %>%
    mutate_at(
      vars(
        income, trust_gov, trust_parliament,
        trust_police, trust_courts, work, sex
      ),
      delete_na_asian
    ) %>% # NAs deleten
    mutate(
      trust_gov = 5 - trust_gov,
      trust_parliament = 5 - trust_parliament,
      trust_police = 5 - trust_police,
      trust_courts = 5 - trust_courts,
      sex = sex - 1, # sex (0/1) codieren
      income = Recode(
        4 - income,
        "8 = NA;
         9 = NA"
      ),
      work = Recode(
        work, # work (0/1) codieren
        "2 = 0"
      ),
      age = Recode(
        se3_2, # missing values l?schen
        "-1 = NA"
      ),
      # demtoday = Recode(
      #   q94,
      #   "-1 = NA;
      #    97 = NA;
      #    98 = NA;
      #    99 = NA"
      # ),
      educ = Recode(
        educ,
        "-1 = NA;
         98 = NA;
         99 = NA"
      ),
      cntry = to_label(country),
      year = year
    ) %>%
    select(
      cntry, year, age, sex, income, educ, work,  
      trust_gov, trust_parliament, trust_police, trust_courts
    )
save(taiwan, file = "codebooks/singles/taiwan.Rdata")

Merging Data

asian_3 %<>%
    mutate_at(
      vars(
        income, educ,
        trust_gov, trust_parliament,
        trust_police, trust_courts
      ),
      range01
    )
myanmar %<>%
    mutate_at(
      vars(
        income, educ,
        trust_gov, trust_parliament,
        trust_police, trust_courts
      ),
      range01
    )
mongolia %<>%
    mutate_at(
      vars(
        income, educ,
        trust_gov, trust_parliament,
        trust_police, trust_courts
      ),
      range01
    )
philip %<>%
    mutate_at(
      vars(
        income, educ,
        trust_gov, trust_parliament,
        trust_police, trust_courts
      ),
      range01
    )  
taiwan %<>%
    mutate_at(
      vars(
        income, educ,
        trust_gov, trust_parliament,
        trust_police, trust_courts
      ),
      range01
    )  
thai %<>%
    mutate_at(
      vars(
        income, educ,
        trust_gov, trust_parliament,
        trust_police, trust_courts
      ),
      range01
    )
malay %<>%
    mutate_at(
      vars(
        income, educ,
        trust_gov, trust_parliament,
        trust_police, trust_courts
      ),
      range01
    )
singapore %<>%
    mutate_at(
      vars(
        income, educ,
        trust_gov, trust_parliament,
        trust_police, trust_courts
      ),
      range01
    )
sk %<>%
    mutate_at(
      vars(
        income, educ,
        trust_gov, trust_parliament,
        trust_police, trust_courts
      ),
      range01
    )
cambodia %<>%
    mutate_at(
      vars(
        income, educ,
        trust_gov, trust_parliament,
        trust_police, trust_courts
      ),
      range01
    )
asian <- rbind(
  asian_3, myanmar, cambodia, sk, singapore, malay, thai, taiwan, philip,
  cambodia, mongolia
) %>% as.tbl()
asian

European Social Survey

load(ESS_raw_url)
ESS <- ESS_raw %>%
  rename(
    # demtoday = dmcntov,
    age = agea,
    year = inwyys,
    trust_gov = trstplt,
    trust_parliament = trstprl,
    trust_police = trstplc,
    trust_courts = trstlgl
  ) %>%
  mutate(
    income = 5 - hincfel,
    educ = Recode(
      eisced,
      "55 = NA"
    ),
    work = ifelse(mnactic == 1, 1, 0),
    sex = gndr - 1, # sex variable erstellen
    cntry = to_label(cntry)
  ) %>%  
  select(
    cntry, year, age, sex, income, educ, work, trust_gov, trust_parliament,
    trust_police, trust_courts
  )
ESS

save(ESS, file = "codebooks/singles/ESS.Rdata")

Merging Everything

afro %<>% 
  mutate(survey = rep("afro", nrow(afro))) %>% 
  mutate(cntry = countrycode(cntry, "country.name", "country.name")) 
latino %<>%  
  mutate(survey = rep("latino", nrow(latino))) %>% 
  mutate(cntry = countrycode(cntry, "country.name", "country.name"))
americas %<>%  
  mutate(survey = rep("americas", nrow(americas))) %>% 
  mutate(cntry = countrycode(cntry, "country.name", "country.name"))
asian %<>%  
  mutate(survey = rep("asian", nrow(asian))) %>% 
  mutate(cntry = countrycode(cntry, "country.name", "country.name"))
wvs %<>%  
  mutate(survey = rep("wvs", nrow(wvs))) %>% 
  mutate(cntry = countrycode(cntry, "country.name", "country.name")) %>%
    mutate_at(
      vars(
        income, educ,
        trust_gov, trust_parliament,
        trust_police, trust_courts
      ),
      range01
    )
ESS %<>%  
  mutate(survey = rep("ESS", nrow(ESS))) %>% 
  mutate(cntry = countrycode(cntry, "country.name", "country.name")) %>%
    mutate_at(
    vars(
      income, educ,
      trust_gov, trust_parliament,
      trust_police, trust_courts
    ),
    stdz
  ) %>%
    mutate_at(
      vars(
        income, educ,
        trust_gov, trust_parliament,
        trust_police, trust_courts
      ),
      range01
    ) 
unique(ESS$cntry)

 [1] "Albania"        "Belgium"        "Bulgaria"       "Switzerland"    "Cyprus"         "Czechia"       
 [7] "Germany"        "Denmark"        "Estonia"        "Spain"          "Finland"        "France"        
[13] "United Kingdom" "Hungary"        "Ireland"        "Israel"         "Iceland"        "Italy"         
[19] "Lithuania"      "Netherlands"    "Norway"         "Poland"         "Portugal"       "Russia"        
[25] "Sweden"         "Slovenia"       "Slovakia"       "Ukraine"        "Kosovo"        

unique(asian$cntry)

 [1] "Japan"               "Hong Kong SAR China" "South Korea"         "China"              
 [5] "Mongolia"            "Philippines"         "Taiwan"              "Thailand"           
 [9] "Indonesia"           "Singapore"           "Vietnam"             "Cambodia"           
[13] "Malaysia"            "Myanmar (Burma)"    

unique(wvs$cntry)

 [1] "Algeria"                 "Argentina"               "Armenia"                 "Australia"              
 [5] "Azerbaijan"              "Bahrain"                 "Belarus"                 "Brazil"                 
 [9] "Colombia"                "Cyprus"                  "Chile"                   "China"                  
[13] "Ecuador"                 "Egypt"                   "Estonia"                 "Georgia"                
[17] "Germany"                 "Ghana"                   "Hong Kong SAR China"     "India"                  
[21] "Iraq"                    "Japan"                   "Jordan"                  "Kazakhstan"             
[25] "Kuwait"                  "Kyrgyzstan"              "Lebanon"                 "Libya"                  
[29] "Malaysia"                "Mexico"                  "Morocco"                 "Netherlands"            
[33] "New Zealand"             "Nigeria"                 "Pakistan"                "Palestinian Territories"
[37] "Peru"                    "Philippines"             "Poland"                  "Qatar"                  
[41] "Romania"                 "Russia"                  "Rwanda"                  "Singapore"              
[45] "Slovenia"                "South Korea"             "South Africa"            "Spain"                  
[49] "Sweden"                  "Taiwan"                  "Thailand"                "Trinidad & Tobago"      
[53] "Tunisia"                 "Turkey"                  "Ukraine"                 "United States"          
[57] "Uruguay"                 "Uzbekistan"              "Yemen"                   "Zimbabwe"               

unique(latino$cntry)

 [1] "Argentina"          "Bolivia"            "Brazil"             "Colombia"           "Costa Rica"        
 [6] "Chile"              "Ecuador"            "El Salvador"        "Spain"              "Guatemala"         
[11] "Honduras"           "Mexico"             "Nicaragua"          "Panama"             "Paraguay"          
[16] "Peru"               "Dominican Republic" "Uruguay"            "Venezuela"         

unique(americas$cntry)

 [1] "Bahamas"            "Barbados"           "Belize"             "Brazil"             "Colombia"          
 [6] "Costa Rica"         "Dominican Republic" "Ecuador"            "El Salvador"        "Guatemala"         
[11] "Guyana"             "Haiti"              "Honduras"           "Jamaica"            "Mexico"            
[16] "Nicaragua"          "Panama"             "Paraguay"           "Peru"               "Suriname"          
[21] "Trinidad & Tobago"  "Uruguay"            "Bolivia"            "Canada"            

unique(afro$cntry)

 [1] "Algeria"             "Burundi"             "Benin"               "Burkina Faso"       
 [5] "Botswana"            "Cameroon"            "Côte d’Ivoire"       "Cape Verde"         
 [9] "Egypt"               "Ghana"               "Guinea"              "Kenya"              
[13] "Lesotho"             "Liberia"             "Madagascar"          "Mauritius"          
[17] "Mali"                "Malawi"              "Mozambique"          "Morocco"            
[21] "Namibia"             "Niger"               "Nigeria"             "South Africa"       
[25] "Senegal"             "Sierra Leone"        "Sudan"               "Swaziland"          
[29] "Tanzania"            "Togo"                "Tunisia"             "Uganda"             
[33] "Zambia"              "Zimbabwe"            "Gabon"               "São Tomé & Príncipe"

merged <- rbind(wvs, latino, afro, americas, asian, ESS)
merged %<>%
  # create dummies
  mutate(
    wvs = ifelse(survey == "wvs", 1, 0),
    afro = ifelse(survey == "afro", 1, 0),
    latino = ifelse(survey == "latino", 1, 0),
    americas = ifelse(survey == "americas", 1, 0),
    asian = ifelse(survey == "asian", 1, 0),
    ESS = ifelse(survey == "ESS", 1, 0)
  ) %>%
  # filter bad countries
  filter(!(cntry == "Egypt" & year == 2013)) %>% # exclude Egypt 2013
  filter(!(cntry == "Libya" & year == 2014)) %>% # exclude Libya 2014
  filter(!(cntry == "Mali" & year == 2012)) %>% # exclude Mali 2012
  filter(!(cntry == "Yemen" & year == 2012)) %>% # exclude Yemen 2012
  filter(!(cntry == "Palestine, State of" & year == 2013)) # exclude Palestine 2013
# adding weight
not_character <- function(x) {
  !(is.character(x))
}
merged %<>%
  group_by(cntry) %>%
  tally() %>%
  mutate(weight = 1000 / n) %>%
  select(cntry, weight) %>%
  left_join(merged, "cntry") %>% 
  mutate_if(not_character, as.numeric)
# select(cntry, year) %>%
# unique %>%
# View
merged

table(merged$wvs)

     0      1 
306779  82618 

table(merged$afro)

     0      1 
286241 103156 

table(merged$latino)

     0      1 
346484  42913 

table(merged$americas)

     0      1 
315547  73850 

table(merged$asian)

     0      1 
357210  32187 

table(merged$ESS)

     0      1 
334724  54673 

Factor Analysis

#TODO todor::
merged %>% 
  select(trust_gov, trust_parliament, 
           trust_police, trust_courts) %>% 
  psych::alpha()

Reliability analysis   
Call: psych::alpha(x = .)

  raw_alpha std.alpha G6(smc) average_r S/N    ase mean   sd
      0.81      0.81    0.78      0.52 4.3 0.0005 0.49 0.26

 lower alpha upper     95% confidence boundaries
0.81 0.81 0.81 

 Reliability if an item is dropped:
                 raw_alpha std.alpha G6(smc) average_r S/N alpha se
trust_gov             0.77      0.77    0.69      0.52 3.3  0.00065
trust_parliament      0.74      0.74    0.67      0.49 2.9  0.00072
trust_police          0.79      0.79    0.73      0.56 3.8  0.00057
trust_courts          0.75      0.75    0.69      0.50 3.0  0.00070

 Item statistics 
                      n raw.r std.r r.cor r.drop mean   sd
trust_gov        364295  0.81  0.80  0.71   0.63 0.49 0.34
trust_parliament 368645  0.82  0.82  0.75   0.67 0.45 0.32
trust_police     382631  0.77  0.76  0.63   0.57 0.52 0.33
trust_courts     372752  0.82  0.82  0.73   0.66 0.50 0.32

merged %>% 
  select(trust_gov, trust_parliament, 
           trust_police, trust_courts) %>% 
  psych::KMO()

Kaiser-Meyer-Olkin factor adequacy
Call: psych::KMO(r = .)
Overall MSA =  0.76
MSA for each item = 
       trust_gov trust_parliament     trust_police     trust_courts 
            0.75             0.74             0.79             0.77 

merged %<>% 
  select(trust_gov, trust_parliament, 
           trust_police, trust_courts) %>% 
  mutate_all(as.numeric) %>% 
  psych::fa(weight = merged$weight) %>% 
  psych::predict.psych(data = merged %>% 
  select(trust_gov, trust_parliament, 
           trust_police, trust_courts)) %>%
  as.data.frame() %>%
  transmute(gov_trust = range01(MR1)) %>%
  cbind(merged)
merged

Some Recoding

Level 2 Data

V-Dem

load(vdems_start_url)
vdems_sub <- vdems_start %>% filter(year %in% 2000:2010)
#tibble(id = 1:1896)
#table(vdems_sub$country_name)
vdem <- vdems_start %>%
  filter(year %in% 2000:2010) %>% 
  group_by(country_name) %>%
  tally %>%
  mutate(cntry = unique(country_name)) %>%
  #DCI Variables
  mutate(delib10 = vdems_sub %>%
           dcast(country_name ~ year, 
                 value.var=c("v2xdl_delib")) %>%
           select(`2000`:`2010`) %>%
           rowMeans) %>%
  mutate(consult10 = vdems_sub %>% 
           dcast(country_name ~ year, 
                 value.var=c("v2dlconslt")) %>%
           select(`2000`:`2010`) %>%
           rowMeans) %>%
  mutate(reason10 = vdems_sub %>% 
           dcast(country_name ~ year, 
                 value.var=c("v2dlreason")) %>%
           select(`2000`:`2010`) %>%
           rowMeans) %>%
  mutate(common10 = vdems_sub %>% 
           dcast(country_name ~ year, 
                 value.var=c("v2dlcommon")) %>%
           select(`2000`:`2010`) %>%
           rowMeans) %>%
  mutate(countr10 = vdems_sub %>% 
           dcast(country_name ~ year, 
                 value.var=c("v2dlcountr")) %>%
           select(`2000`:`2010`) %>%
           rowMeans) %>%
  mutate(engage10 = vdems_sub %>% 
           dcast(country_name ~ year, 
                 value.var=c("v2dlengage")) %>%
           select(`2000`:`2010`) %>%
           rowMeans) %>%
  mutate(delibdem10 = vdems_sub %>% 
           dcast(country_name ~ year, 
                 value.var=c("v2x_delibdem")) %>%
           select(`2000`:`2010`) %>%
           rowMeans) %>%
  # Control Variables
  mutate(polity10 = vdems_sub %>% 
           dcast(country_name ~ year, 
                 value.var=c("e_fh_ipolity2")) %>%
           select(`2000`:`2010`) %>%
           rowMeans) %>%
  mutate(poly10 = vdems_sub %>% 
           dcast(country_name ~ year, 
                 value.var=c("v2x_polyarchy")) %>%
           select(`2000`:`2010`) %>%
           rowMeans) %>%
  mutate(rol10 = vdems_sub %>% 
           dcast(country_name ~ year, 
                 value.var=c("e_wbgi_rle")) %>%
           select(`2000`:`2010`) %>%
           rowMeans) %>%
#  mutate(gdp10 = vdems_sub %>% filter(year %in% 2000:2008) %>% 
#           dcast(country_name ~ year, 
#                 value.var=c("e_GDP_Per_Cap_Haber_Men_2")) %>% 
#           select(`2000`:`2010`) %>%
#           rowMeans) %>%
  mutate(riw10 = vdems_sub %>% 
           dcast(country_name ~ year, 
                 value.var=c("e_v2x_regime_ci")) %>%
           select(`2000`:`2010`) %>%
           rowMeans) %>%
  mutate(corecivil10 = vdems_sub %>% 
           dcast(country_name ~ year, 
                 value.var=c("v2xcs_ccsi")) %>%
           select(`2000`:`2010`) %>%
           rowMeans) %>%
  mutate(pop10 = vdems_sub %>% 
           dcast(country_name ~ year, 
                 value.var=c("e_mipopula")) %>%
           select(`2000`:`2010`) %>%
           rowMeans) %>%
  mutate(corruption10 = vdems_sub %>% 
           dcast(country_name ~ year, 
                 value.var=c("v2x_corr")) %>%
           select(`2000`:`2010`) %>%
           rowMeans) %>%
  mutate(polkill10 = vdems_sub %>% 
           dcast(country_name ~ year, 
                 value.var=c("v2x_clphy")) %>%
           select(`2000`:`2010`) %>%
           rowMeans) %>%
  mutate(educ10 = vdems_sub %>% 
           dcast(country_name ~ year, 
                 value.var=c("e_peaveduc")) %>%
           select(`2000`:`2010`) %>%
           rowMeans) %>%
  mutate(equal_distrib10 = vdems_sub %>% 
           dcast(country_name ~ year, 
                 value.var=c("v2xeg_eqdr")) %>%
           select(`2000`:`2010`) %>%
           rowMeans) %>%  
   mutate(div_party10 = vdems_sub %>% 
           dcast(country_name ~ year, 
                 value.var=c("v2x_divparctrl")) %>%
           select(`2000`:`2010`) %>%
           rowMeans) %>%  
   mutate(div_power10 = vdems_sub %>% 
           dcast(country_name ~ year, 
                 value.var=c("v2x_feduni")) %>%
           select(`2000`:`2010`) %>%
           rowMeans) %>%  
   mutate(accountability10 = vdems_sub %>% 
           dcast(country_name ~ year, 
                 value.var=c("v2x_accountability")) %>%
           select(`2000`:`2010`) %>%
           rowMeans) %>%    
   mutate(directdem10 = vdems_sub %>% 
           dcast(country_name ~ year, 
                 value.var=c("v2xdd_dd")) %>%
           select(`2000`:`2010`) %>%
           rowMeans) %>%      
   mutate(eq_educ10 = vdems_sub %>% 
           dcast(country_name ~ year, 
                 value.var=c("v2peedueq")) %>%
           select(`2000`:`2010`) %>%
           rowMeans) %>%       
    mutate(eq_health10 = vdems_sub %>% 
           dcast(country_name ~ year, 
                 value.var=c("v2pehealth")) %>%
           select(`2000`:`2010`) %>%
           rowMeans) %>%  
     mutate(controlcorruption10 = vdems_sub %>% 
           dcast(country_name ~ year, 
                 value.var=c("e_wbgi_cce")) %>%
           select(`2000`:`2010`) %>%
           rowMeans) %>%  
     mutate(eff10 = vdems_sub %>% 
           dcast(country_name ~ year, 
                 value.var=c("e_wbgi_gee")) %>%
           select(`2000`:`2010`) %>%
           rowMeans) %>% 
     mutate(stability10 = vdems_sub %>% 
           dcast(country_name ~ year, 
                 value.var=c("e_wbgi_pse")) %>%
           select(`2000`:`2010`) %>%
           rowMeans) %>%   
     mutate(unified_dem10 = vdems_sub %>% 
           dcast(country_name ~ year, 
                 value.var=c("e_uds_mean")) %>%
           select(`2000`:`2010`) %>%
           rowMeans) %>%     
     mutate(gdp_growth10 = vdems_sub %>% 
           dcast(country_name ~ year, 
                 value.var=c("e_migdpgrolns")) %>%
           select(`2000`:`2010`) %>%
           rowMeans) %>%       
     mutate(infla10 = vdems_sub %>% 
           dcast(country_name ~ year, 
                 value.var=c("e_miinflat")) %>%
           select(`2000`:`2010`) %>%
           rowMeans) %>%       
     # mutate(lifeexp10 = vdems_sub %>% 
     #       dcast(country_name ~ year, 
     #             value.var=c("e_pelifeex")) %>%
     #       select(`2000`:`2010`) %>%
     #       rowMeans) %>%        
    mutate(urbanratio10 = vdems_sub %>% 
           dcast(country_name ~ year, 
                 value.var=c("e_miurbani")) %>%
           select(`2000`:`2010`) %>%
           rowMeans) %>%        
  
#  mutate(gini10 = vdems_sub %>% 
#           dcast(country_name ~ year, 
#                 value.var=c("e_peginiwi")) %>%
#           select(`2000`:`2010`) %>%
#           rowMeans) %>% 
  # Dummy variables
  mutate(pol_round = round(polity10 * 2 - 10)) %>% 
  mutate(polity_demdummy = ifelse(pol_round > 5, 1, 0)) %>% 
  mutate(polity_anodummy = ifelse(pol_round >= -5 & pol_round <= 5, 1, 0)) %>% 
  mutate(polity_autodummy = ifelse(pol_round < -5, 1, 0)) %>% 
  mutate(regime = case_when(
    polity_autodummy == 1 ~ "auto",
    polity_anodummy == 1 ~ "ano",
    polity_demdummy == 1 ~ "demo"
    )
  ) %>%
  mutate(regime = factor(regime, levels = c("demo", "ano", "auto"))) %>% 
  mutate(cntry = ifelse(str_detect(cntry, "ietnam"), "Vietnam", cntry)) %>%   
  mutate(cntry = countrycode(cntry,"country.name","country.name")) %>% 
  select(-n)
# vdem$cntry[40] <- "Viet Nam"
# ifelse(round(vdem$polity10 * 2 -10) >= -5 & round(vdem$polity10 * 2 -10) <= 5, 1, 0)
table(vdem$regime)

demo  ano auto 
  72   77   18 

print(levels(vdem$regime)) 

[1] "demo" "ano"  "auto"

# DURATION OF REGIME?! e_democracy_duration

QoG

load("data/qog.Rdata")
count_na <- function(x) sum(is.na(x))
qog10 <- qog %>% 
  filter(year %in% 2000:2010) %>% 
  mutate(count_na = apply(., 1, count_na)) %>% 
  filter(count_na < 2050) %>%  
  mutate(cntry = countrycode(ccodealp, "iso3c","country.name"))

Some values were not matched unambiguously: SCG

qog10 <- qog %>% 
  filter(year %in% 2000:2010) %>% 
  mutate(count_na = apply(., 1, count_na)) %>% 
  filter(count_na < 2050) %>%  
  mutate(cntry = countrycode(ccodealp, "iso3c","country.name")) %>% 
  group_by(cntry) %>%
  tally() %>% 
  mutate(ethnic10 = qog10 %>%
           dcast(cntry ~ year, 
                 value.var = c("al_ethnic"), 
          fun.aggregate = mean) %>%
           select(`2000`:`2010`) %>%
           rowMeans) %>% 
  mutate(gdp10 = qog10 %>%
           dcast(cntry ~ year, 
                 value.var = c("gle_cgdpc"), 
          fun.aggregate = mean) %>%
           select(`2000`:`2010`) %>%
           rowMeans) %>% 
  mutate(corrupt10 = qog10 %>%
           dcast(cntry ~ year, 
                 value.var = c("bci_bci"), 
          fun.aggregate = mean) %>%
           select(`2000`:`2010`) %>%
           rowMeans) %>% 
  mutate(lifeexp10 = qog10 %>%
           dcast(cntry ~ year, 
                 value.var = c("ihme_lebs0001"), 
          fun.aggregate = mean) %>%
           select(`2000`:`2010`) %>%
           rowMeans) %>% 
  mutate(mortinf10 = qog10 %>%
           dcast(cntry ~ year, 
                 value.var = c("wdi_mortinf"), 
          fun.aggregate = mean) %>%
           select(`2000`:`2010`) %>%
           rowMeans) %>% 
  mutate(urbanpop10 = qog10 %>%
           dcast(cntry ~ year,
                 value.var = c("wdi_popurb"),
          fun.aggregate = mean) %>%
           select(`2000`:`2010`) %>%
           rowMeans) %>%
  mutate(qogpop10 = qog10 %>%
           dcast(cntry ~ year,
                 value.var = c("pwt_pop"),
          fun.aggregate = mean) %>%
           select(`2000`:`2010`) %>%
           rowMeans) %>%  
  # mutate(work10 = qog10 %>%
  #          dcast(cntry ~ year, 
  #                value.var = c("wdi_unemp"), 
  #         fun.aggregate = mean) %>%
  #          select(`2000`:`2010`) %>%
  #          rowMeans) %>% 
  mutate(frac_elites10 =  qog10 %>% filter(year %in% 2004:2010) %>%
           dcast(cntry ~ year, 
                 value.var = c("ffp_fe"), 
          fun.aggregate = mean) %>%
           select(`2004`:`2010`) %>%
           rowMeans) %>% 
  mutate(glob10 = qog10 %>%
           dcast(cntry ~ year, 
                 value.var = c("dr_ig"), 
          fun.aggregate = mean) %>%
           select(`2000`:`2010`) %>%
           rowMeans) %>% 
  select(-n)

Some values were not matched unambiguously: SCG

#  mutate(al_ethnic = as.numeric(al_ethnic)) %>% 
  
qog10

# qog10 %>%
#   # select(cntry, cname, ccodecow) %>%
#   group_by(cntry) %>%
#   slice(1) %>%
#   arrange(cntry)
# # 
# # countrycode()
# 
# qog %>% 
#   filter(year %in% 2000:2010) %>% 
#   filter(str_detect(cname, "Germany")) %>% 
#   select(cname, al_ethnic, gle_cgdpc)
# 
# 
qog %>%
  filter(year %in% 2000:2010) %>%
  filter(str_detect(cname, "Taiwan")) %>%
  select(cname, wdi_lifexp)

# options(scipen=999)

Gini Coefficient

load("data/swiid6_1.rda")
gini <- swiid_summary
gini_sub <- gini %>% filter(year %in% 2000:2010)
gini %<>% 
  filter(year %in% 2000:2010) %>% 
  group_by(country) %>%
  tally %>%
  mutate(cntry = unique(country)) %>%
  #DCI Variables
  mutate(gini10 = gini_sub %>%
           dcast(country ~ year, 
                 value.var = c("gini_disp")) %>%
           select(`2000`:`2010`) %>%
           rowMeans) %>% 
  mutate(cntry = countrycode(cntry,"country.name","country.name")) %>%  
  select(cntry, gini10)

Some values were not matched unambiguously: Micronesia

Merging Time

Level 2

level2 <- merge(x = qog10, y = vdem, by = "cntry") %>% merge(gini, by = "cntry")
level2

Ind + Country

# load("data/level1.Rdata")
combined <- merge(x = merged, y = level2, by = "cntry") %>% 
              mutate(cntryears = paste(cntry, year))
#combined$cntryears <- paste(combined$cntry,combined$year)
vdem2 <- vdems_start %>%
  filter(year %in% 2010:2015) %>% 
  mutate(country_name = ifelse(str_detect(country_name, "ietnam"), "Vietnam", country_name)) %>%   
  mutate(cntry = countrycode(country_name,"country.name","country.name")) #%>% 
 # select(cntry, country_name)
# vdem2$cntry[996] <- "Viet Nam"
# vdem2$cntry[997] <- "Viet Nam"
# vdem2$cntry[998] <- "Viet Nam"
# vdem2$cntry[999] <- "Viet Nam"
# vdem2$cntry[1000] <- "Viet Nam"
# vdem2$cntry[1001] <- "Viet Nam"
# v2x_freexp_thick freedom of expression
combined <- vdem2 %>% 
  mutate(cntryears = paste(cntry, year)) %>% 
  mutate(discuss_unsel = v2xcl_disc) %>% 
  select(cntryears, discuss_unsel) %>% 
  merge(combined, by = "cntryears") %>% 
                  plyr::ddply(~cntry,
                        summarise, 
                        discuss = mean(discuss_unsel, na.rm=T)) %>% 
  merge(combined, by = "cntry")
save(combined,file = "data/combined.Rdata")
save(level2,file = "data/level2.Rdata")

Weighting, Remove NAs and Normalizing

Individual Level

Macro Level

macro  <- combined %>% 
                # mutate(gov_trust = range01(gov_trust)) %>% 
                # mutate(demtoday = range01(demtoday)*100) %>% 
                group_by(cntry) %>% 
                dplyr::summarise(
                  mean_gov = mean(gov_trust, na.rm=T),
                  mean_gov_low = mean(gov_trust_low, na.rm=T),
                  mean_gov_high = mean(gov_trust_high, na.rm=T),
                  discuss = mean(discuss, na.rm=T),
                  ) %>% 
                merge(level2, by = "cntry") %>% 
                zap_inf() %>% 
                mutate(urbanratio10 = ifelse(is.na(urbanratio10), urbanpop10, urbanratio10)) %>% 
                mutate(pop10 = ifelse(is.na(pop10), qogpop10 * 1000000, pop10)) %>% 
                select(cntry, mean_gov, mean_gov_low, mean_gov_high, 
                       delib10:polity10, 
                       pol_round, polity_demdummy, polity_anodummy, 
                       polity_autodummy, polity_autodummy,
                       regime, discuss, #discuss_round, 
                       gdp10, corrupt10, lifeexp10, pop10, 
                       poly10, riw10, unified_dem10, urbanratio10) %>% 
                drop_na(mean_gov) %>% 
                mutate(count_na = apply(., 1, count_na)) %>% 
  mutate_at(vars(delib10:polity10, polity_demdummy:polity_autodummy, discuss:urbanratio10), range01)
macro_dem <- macro %>% 
                      filter(polity_demdummy == 1) %>% 
  mutate_at(vars(delib10:polity10, polity_demdummy:polity_autodummy, discuss:urbanratio10), range01)
macro_aut <- macro %>% 
                      filter(polity_demdummy == 0) %>% 
  mutate_at(vars(delib10:polity10, polity_demdummy:polity_autodummy, discuss:urbanratio10), range01)
# table_stuff2<-subset(table_stuff,
#                      !is.na(table_stuff$regime) &
#                        table_stuff$cntry!="Qatar" &
#                       table_stuff$cntry!="Uzbekistan")
save(macro, file = "data/macro.Rdata")
save(macro_dem, file = "data/macro_dem.Rdata")
save(macro_aut, file = "data/macro_aut.Rdata")

Crap

SEM Index

merged2 <- merged %>%
  mutate(gov_trust = trust_gov + trust_parliament + 
           trust_police + trust_courts) %>%
  filter(!is.na(gov_trust))

merged3 <- merged %>%
  mutate(gov_trust = trust_gov + trust_parliament + 
           trust_police + trust_courts) %>%
  filter(is.na(gov_trust))

svy.df <- survey::svydesign(id= ~1,
                              weights= ~weight,
                              data= merged) 

model <- '# measurement model 1
gov_trust2 =~ 1*trust_gov + trust_parliament + 
trust_police + trust_courts
trust_gov ~~ trust_parliament
'

merged <- merged %>%
  mutate_at(vars(trust_gov, trust_parliament,
                 trust_police, trust_courts), as.numeric)

# cor(na.omit(data.frame(merged$trust_gov,
#               merged$trust_police,
#               merged$trust_courts,
#               merged$trust_parliament,
#               merged$demtoday)))

lavaan_model1<-cfa(model, meanstructure = T, 
                   data = as.data.frame(merged),
                   estimator= "MLM")

fit_a1<-lavaan.survey(lavaan_model1, 
            estimator= "MLM", survey.design=svy.df)
summary(fit_a1, standardized=TRUE,fit.measures = TRUE, rsq = T)

merged4<-cbind(merged2, predict(fit_a1, newdata = merged2))
merged4$gov_trust2<-range01(merged4$gov_trust2)

# head(merged4)

merged<-plyr::rbind.fill(merged3,merged4)
merged$gov_trust<-merged$gov_trust2
head(merged)

compare_cntry<-data.frame(table(merged$cntry,merged$year))
compare_cntry<-tidyr::spread(compare_cntry,Var2,Freq)
compare_cntry[compare_cntry==0]<-100000
#compare_cntry<-na.omit(compare_cntry)
indie<-as.numeric(apply(compare_cntry[,-1],1,sum))
compare_cntry$double<-indie<900000
compare_cntry[compare_cntry==100000]<-0
compare_cntry2 <- subset(compare_cntry,compare_cntry$double==TRUE)

compare_cntry3 <- subset(merged,merged$cntry %in% compare_cntry2$Var1)

y1<-subset(compare_cntry3,compare_cntry3$cntry=="Belize" &
                      compare_cntry3$year==2012)$gov_trust

y2<-subset(compare_cntry3,compare_cntry3$cntry=="Belize" &
         compare_cntry3$year==2014)$gov_trust

t.test(y1,y2)

y1<-subset(compare_cntry3,compare_cntry3$cntry=="Bolivia (Plurinational State of)" &
             compare_cntry3$year==2012)$gov_trust

y2<-subset(compare_cntry3,compare_cntry3$cntry=="Bolivia (Plurinational State of)" &
             compare_cntry3$year==2013)$gov_trust

t.test(y1,y2)

y1<-subset(compare_cntry3,compare_cntry3$cntry=="South Africa" &
             compare_cntry3$year==2011)$gov_trust

y2<-subset(compare_cntry3,compare_cntry3$cntry=="South Africa" &
             compare_cntry3$year==2013)$gov_trust

t.test(y1,y2)

       
subset(merged,merged$year==2012)

library(dplyr)
compare_cntry4 <- compare_cntry3 %>%
  group_by(cntry,year) %>%
  summarise_all(funs(mean(., na.rm=TRUE)))

compare_cntry4$cntryear<-paste(compare_cntry4$cntry,compare_cntry4$year)
  
as.data.frame(compare_cntry4[,c(17,8)])
  
?ddply
head(afro)

afrocntry<-unique(afro$cntry)
latinocntry<-unique(latino2013$cntry)
arabcntry<-unique(arab3$cntry)
wvscntry<-unique(wvs$cntry)

cntry_table<- data.frame(as.character(wvscntry),
                         c(as.character(arabcntry),rep("-",48)),
                         c(as.character(latinocntry),rep("-",41)),
                         c(as.character(afrocntry),rep("-",24)))

colnames(cntry_table) <- c("wvs","arab","latino","afro")

arrange(cntry_table, wvs, afro)



table(combined$politcat.x)
mplusdata <- combined %>% dplyr::select(cntry, delib10, 
                                        trust_gov, trust_parliament, 
                                        trust_police,trust_courts,weight)
mplusdata <- na.omit(mplusdata)
mplusdata$cntry <- as.numeric(mplusdata$cntry)
write_csv(mplusdata, path = "mplusdata.csv", col_names = F)

mplusdata2 <- combined %>% dplyr::select(cntry, delib10, 
                                        gov_trust,weight)
mplusdata2 <- na.omit(mplusdata2)
mplusdata2$cntry <- as.numeric(mplusdata2$cntry)
write_csv(mplusdata2, path = "mplusdat2a.csv", col_names = F)

mplusdata2_dem <- combined_dem %>% dplyr::select(cntry, delib10, 
                                         gov_trust,weight)
mplusdata2_dem <- na.omit(mplusdata2_dem)
mplusdata2_dem$cntry <- as.numeric(mplusdata2_dem$cntry)
write_csv(mplusdata2_dem, path = "mplusdat2_dema.csv", col_names = F)

mplusdata2_aut <- combined_aut %>% dplyr::select(cntry, delib10, 
                                         gov_trust,weight)
mplusdata2_aut <- na.omit(mplusdata2_aut)
mplusdata2_aut$cntry <- as.numeric(mplusdata2_aut$cntry)
write_csv(mplusdata2_aut, path = "mplusdat2_auta.csv", col_names = F)

combined <- merge(combined, physi2_s, by = "cntry")

combined_dem <- subset(combined,combined$polity_demdummy==1)

mplusdata_dem <- combined_dem %>% dplyr::select(cntry, delib10, 
                                        trust_gov, trust_parliament, 
                                        trust_police,trust_courts,weight)
mplusdata_dem <- na.omit(mplusdata_dem)
mplusdata_dem$cntry <- as.numeric(mplusdata_dem$cntry)
write_csv(mplusdata_dem, path = "mplusdata_dem.csv", col_names = F)


combined_aut <- subset(combined,combined$polity_demdummy==0)

mplusdata_aut <- combined_aut %>% dplyr::select(cntry, delib10, 
                                                trust_gov, trust_parliament, 
                                                trust_police,trust_courts,weight)
mplusdata_aut <- na.omit(mplusdata_aut)
mplusdata_aut$cntry <- as.numeric(mplusdata_aut$cntry)
write_csv(mplusdata_aut, path = "mplusdata_aut.csv", col_names = F)


#data_wids <- dcast(merged, cntry~year, 
#                    value.var=c("year"))
#data_wids2 <- as.data.frame(lapply(data_wids[,-1],function(n) 0<n))
#data_wids3 <- as.data.frame(apply(data_wids2,2,as.numeric))
#data_wids3 <- cbind(data_wids[,1],data_wids3)
#names(data_wids3)[1]<-"cntry"
#table_stuff <- merge(data_wids3 ,table_stuff, by="cntry")

#table_stuff$polity10[table_stuff$cntry=="Tunisia"] <- 5.32382
#table_stuff2$polity10[table_stuff2$cntry=="Tunisia"] <- 5.32382

head(table_stuff)
cor(na.omit(data.frame(table_stuff$legit,table_stuff$legit2,table_stuff$legit3,
                       table_stuff$mean_gov)))

table_stuff$gni <- table_stuff$gni_c

table_stuff$gni_c[table_stuff$gni <= 1025] <- "low"
table_stuff$gni_c[table_stuff$gni > 1026 & table_stuff$gni <= 4035] <- "lower-middle"
table_stuff$gni_c[table_stuff$gni > 4036 & table_stuff$gni <= 12475] <- "upper-middle"
table_stuff$gni_c[table_stuff$gni > 12475] <- "high"

table_stuff$gni_c2 <- table_stuff$gni_c

table_stuff$gni_c2[table_stuff$gni_c == "lower-middle"] <- "low"
table_stuff$gni_c2[table_stuff$gni_c == "upper-middle"] <- "high"

table(table_stuff$gni_c)
table(table_stuff$gni_c2)

The cut-off points are HDI of less than 0.550
for low human development, 0.550-0.699 for medium human
development, 0.700-0.799 for high human development and
0.800 or greater for very high human development.

#table_stuff$hdi_c <- table_stuff$hdi

#table_stuff$hdi_c[table_stuff$hdi < 0.550] <- "low"
#table_stuff$hdi_c[table_stuff$hdi >= 0.550 & table_stuff$hdi <= 0.699] <- "medium"
#table_stuff$hdi_c[table_stuff$hdi >= 0.700 & table_stuff$hdi <= 0.899] <- "high"
#table_stuff$hdi_c[table_stuff$hdi >= 0.90] <- "very high"

#table(table_stuff$hdi_c)



table(round(table_stuff$terror))

table_stuff$mean_gov2 <- table_stuff$mean_gov
table_stuff$mean_gov2[is.na(table_stuff$mean_gov2)]<-999

table_stuff$mean_gov3 <- table_stuff$mean_gov
table_stuff$mean_gov3[is.na(table_stuff$mean_gov3)]<-999

table_stuff$mean_gov4 <- table_stuff$mean_gov
table_stuff$mean_gov4[is.na(table_stuff$mean_gov4)]<-999

table_stuff$mean_gov5 <- table_stuff$mean_gov
table_stuff$mean_gov5[is.na(table_stuff$mean_gov5)]<-999

table_stuff$mean_gov6 <- table_stuff$mean_gov
table_stuff$mean_gov6[is.na(table_stuff$mean_gov6)]<-999

table_stuff$mean_gov7 <- table_stuff$mean_gov
table_stuff$mean_gov7[is.na(table_stuff$mean_gov7)]<-999

table_stuff$mean_gov8 <- table_stuff$mean_gov
table_stuff$mean_gov8[is.na(table_stuff$mean_gov8)]<-999

table_stuff$mean_gov9 <- table_stuff$mean_gov
table_stuff$mean_gov9[is.na(table_stuff$mean_gov9)]<-999

table_stuff$physviol2 <- table_stuff$physviol
table_stuff$physviol2[is.na(table_stuff$physviol2)] <- 999

table_stuff$terror2 <- round(table_stuff$terror)
table_stuff$terror2[is.nan(table_stuff$terror2)] <- 999

table_stuff$discuss2 <- round(table_stuff$discuss*4)
table_stuff$discuss2[is.nan(table_stuff$discuss2)] <- 999



table(table_stuff$physviol)
table_stuff$mean_gov2[table_stuff$physviol2<=0.5 & table_stuff$physviol2>0.4] <- table_stuff$mean_gov2[table_stuff$physviol2<=0.5 & table_stuff$physviol2>0.4]-5
table_stuff$mean_gov2[table_stuff$physviol2<=0.4 & table_stuff$physviol2>0.3] <- table_stuff$mean_gov2[table_stuff$physviol2<=0.4 & table_stuff$physviol2>0.3]-10
table_stuff$mean_gov2[table_stuff$physviol2<=0.3 & table_stuff$physviol2>0.2] <- table_stuff$mean_gov2[table_stuff$physviol2<=0.3 & table_stuff$physviol2>0.2]-15
table_stuff$mean_gov2[table_stuff$physviol2<=0.2 & table_stuff$physviol2>0.1] <- table_stuff$mean_gov2[table_stuff$physviol2<=0.2 & table_stuff$physviol2>0.1]-20
table_stuff$mean_gov2[table_stuff$physviol2<=0.1 & table_stuff$physviol2>=0]  <- table_stuff$mean_gov2[table_stuff$physviol2<=0.1 & table_stuff$physviol2>=0]  -25

table_stuff$mean_gov2[table_stuff$mean_gov2>100] <- NA
table(table_stuff$mean_gov2)

table_stuff$mean_gov3[table_stuff$physviol2<=0.5 & table_stuff$physviol2>0.4] <- table_stuff$mean_gov3[table_stuff$physviol2<=0.5 & table_stuff$physviol2>0.4]-8
table_stuff$mean_gov3[table_stuff$physviol2<=0.4 & table_stuff$physviol2>0.3] <- table_stuff$mean_gov3[table_stuff$physviol2<=0.4 & table_stuff$physviol2>0.3]-16
table_stuff$mean_gov3[table_stuff$physviol2<=0.3 & table_stuff$physviol2>0.2] <- table_stuff$mean_gov3[table_stuff$physviol2<=0.3 & table_stuff$physviol2>0.2]-24
table_stuff$mean_gov3[table_stuff$physviol2<=0.2 & table_stuff$physviol2>0.1] <- table_stuff$mean_gov3[table_stuff$physviol2<=0.2 & table_stuff$physviol2>0.1] -32
table_stuff$mean_gov3[table_stuff$physviol2<=0.1 & table_stuff$physviol2>=0] <-  table_stuff$mean_gov3[table_stuff$physviol2<=0.1 & table_stuff$physviol2>=0]  -30

table_stuff$mean_gov3[table_stuff$mean_gov3>100] <- NA
table(table_stuff$mean_gov3)

table_stuff$mean_gov4[table_stuff$terror2==5] <- table_stuff$mean_gov4[table_stuff$terror2==5]-5
table_stuff$mean_gov4[table_stuff$terror2==4] <- table_stuff$mean_gov4[table_stuff$terror2==4]-10
table_stuff$mean_gov4[table_stuff$terror2==3] <- table_stuff$mean_gov4[table_stuff$terror2==3]-15

table_stuff$mean_gov4[table_stuff$mean_gov4>100] <- NA
table(table_stuff$mean_gov4)

table_stuff$mean_gov5[table_stuff$terror2==5] <- table_stuff$mean_gov5[table_stuff$terror2==5]-8
table_stuff$mean_gov5[table_stuff$terror2==4] <- table_stuff$mean_gov5[table_stuff$terror2==4]-16
table_stuff$mean_gov5[table_stuff$terror2==3] <- table_stuff$mean_gov5[table_stuff$terror2==3]-24

table_stuff$mean_gov5[table_stuff$mean_gov5>100] <- NA
table(table_stuff$mean_gov5)

table_stuff$mean_gov6[table_stuff$discuss2==2] <- table_stuff$mean_gov6[table_stuff$discuss2==2] -5
table_stuff$mean_gov6[table_stuff$discuss2==1] <- table_stuff$mean_gov6[table_stuff$discuss2==1] -10

table_stuff$mean_gov6[table_stuff$mean_gov6>100] <- NA
table(table_stuff$mean_gov6)

table_stuff$mean_gov7[table_stuff$discuss2==2] <- table_stuff$mean_gov7[table_stuff$discuss2==2] -10
table_stuff$mean_gov7[table_stuff$discuss2==1] <- table_stuff$mean_gov7[table_stuff$discuss2==1] -20

table_stuff$mean_gov7[table_stuff$mean_gov7>100] <- NA
table(table_stuff$mean_gov7)


load(qog_url)

qog10 <- subset(qog,qog$year==2000 |
                  qog$year==2001 |  
                  qog$year==2002 | 
                  qog$year==2003 | 
                  qog$year==2004 | 
                  qog$year==2005 | 
                  qog$year==2006 | 
                  qog$year==2007 | 
                  qog$year==2008 | 
                  qog$year==2009 |
                  qog$year==2010)

library(countrycode)
qog10$cntry<-countrycode(qog10$ccodecow, "cown","country.name")
qog10$al_ethnic<-as.numeric(qog10$al_ethnic)
#tidyr::gather(qog10,c("cntry","year"),"al_ethnic")

qog10a <- qog10 %>% dplyr::select(cntry,year,al_ethnic) %>%
  as.data.frame()
data_wide17 <- reshape(data = qog10a,
                       idvar = "cntry",
                       v.names = "al_ethnic",
                       timevar = "year",
                       direction = "wide")

#qog10hdi <- qog10 %>% dplyr::select(cntry,year,undp_hdi) %>%
#  as.data.frame()
#data_wide_hdi <- hdi
  #reshape(data = qog10hdi,
                      # idvar = "cntry",
                      # v.names = "undp_hdi",
                      # timevar = "year",
                      # direction = "wide")



table(qog$al_ethnic,qog$year)

qog14 <- subset(qog,qog$year==2014)
qog14$cntry<-countrycode(qog14$ccodecow, "cown","country.name")
qog14 <- qog14 %>% dplyr::select(cntry,year,cspf_legit) %>%
  as.data.frame()
data_wide18 <- reshape(data = qog14,
                       idvar = "cntry",
                       v.names = "cspf_legit",
                       timevar = "year",
                       direction = "wide")

qog15 <- subset(qog,qog$year==2015)
qog15$cntry<-countrycode(qog15$ccodecow, "cown","country.name")
qog15 <- qog15 %>% dplyr::select(cntry,year,ffp_sl) %>%
  as.data.frame()
data_wide19 <- reshape(data = qog15,
                       idvar = "cntry",
                       v.names = "ffp_sl",
                       timevar = "year",
                       direction = "wide")


qog10$ciri_physint<-as.numeric(qog10$ciri_physint)
qog10a <- qog10 %>% dplyr::select(cntry,year,ciri_physint) %>%
  as.data.frame()
data_wide20 <- reshape(data = qog10a,
                       idvar = "cntry",
                       v.names = "ciri_physint",
                       timevar = "year",
                       direction = "wide")

qog10$gd_ptss <-as.numeric(qog10$gd_ptss)
qog10a <- qog10 %>% dplyr::select(cntry,year,gd_ptss) %>%
  as.data.frame()
data_wide21 <- reshape(data = qog10a,
                       idvar = "cntry",
                       v.names = "gd_ptss",
                       timevar = "year",
                       direction = "wide")

qog14<-subset(qog,qog$year==2014)
qog13<-subset(qog,qog$year==2013)
qog12<-subset(qog,qog$year==2012)
qog11<-subset(qog,qog$year==2011)
qog10<-subset(qog,qog$year==2010)
qog8<-subset(qog,qog$year==2008)
qog7<-subset(qog,qog$year==2007)
qog6<-subset(qog,qog$year==2006)
qog5<-subset(qog,qog$year==2005)
qog4<-subset(qog,qog$year==2004)
qog3<-subset(qog,qog$year==2003)
qog2<-subset(qog,qog$year==2002)
qog1<-subset(qog,qog$year==2001)
qog0<-subset(qog,qog$year==2000)



table(is.na(qog$wel_culture),qog$year)
culreg<- pmax(qog14$wel_culture, qog13$wel_culture, na.rm = TRUE)
culreg<- pmax(culreg, qog12$wel_culture, na.rm = TRUE)
culreg<- pmax(culreg, qog11$wel_culture, na.rm = TRUE)
culreg<- pmax(culreg, qog10$wel_culture, na.rm = TRUE)
culreg<- pmax(culreg, qog8$wel_culture, na.rm = TRUE)
culreg<- pmax(culreg, qog7$wel_culture, na.rm = TRUE)
culreg<- pmax(culreg, qog6$wel_culture, na.rm = TRUE)
culreg<- pmax(culreg, qog5$wel_culture, na.rm = TRUE)
culreg<- pmax(culreg, qog4$wel_culture, na.rm = TRUE)
culreg<- pmax(culreg, qog3$wel_culture, na.rm = TRUE)
culreg<- pmax(culreg, qog2$wel_culture, na.rm = TRUE)
culreg<- pmax(culreg, qog1$wel_culture, na.rm = TRUE)
culreg<- pmax(culreg, qog0$wel_culture, na.rm = TRUE)
culreg<- as.character(sjmisc::to_label(culreg))
culreg<-as.data.frame(cbind(culreg,qog14$ccodecow))
oreast  <- subset(culreg, culreg$culreg == "Orthodox East")
oreast

#culreg<- pmax(culreg, qog10$wel_culture, na.rm = TRUE)

qog14$ht_region 

culreg$cntry<-countrycode(culreg$V2, "cown","country.name")
tabeletto<-merge(table_stuff,culreg,by="cntry")
tabeletto<-as.data.frame(cbind(table_stuff$regions,
                               as.character(table_stuff$cntry),
                               as.character(tabeletto$culreg)))
tabeletto <- tabeletto[order(tabeletto$V1, tabeletto$V3),]
#edit(tabeletto)

culreg<-data.frame(cbind(tabeletto,culreg))
culreg<-culreg[,c(2,4)]
names(culreg)[1]<-c("cntry")

table_stuff3 <- merge(table_stuff,culreg,by="cntry")

table(as.character(tabeletto$V3))

qog100 <- subset(qog,qog$year==2010)
qog100$cntry<-countrycode(sjmisc::to_label(qog100$ccodecow), "cown","country.name")
qog100 <- qog100 %>% dplyr::select(cntry,year,ht_regtype1) %>%
  as.data.frame()
qog100$regtype<-sjmisc::to_label(qog100$ht_regtype1)
qog100$year <- NULL
qog100$ht_regtype1 <- NULL
table(qog100$regtype)


ethnic10<-as.numeric(rowMeans(data_wide17[,2:12])) # mean over last 10 years (2000 - 2010)
physint10<-as.numeric(rowMeans(data_wide20[,2:12])) # mean over last 10 years (2000 - 2010)
terror10<-as.numeric(rowMeans(data_wide21[,2:12])) # mean over last 10 years (2000 - 2010)
legit<-as.numeric(data_wide18[,2]) # mean over last 10 years (2000 - 2010)
legit2<-as.numeric(data_wide19[,2]) # mean over last 10 years (2000 - 2010)
qogthingy<-data.frame(data_wide19[,1],ethnic10,legit,legit2,physint10,terror10)
colnames(qogthingy)[1]<-c("cntry")
qogthingy<-merge(x=qogthingy, y=qog100, by="cntry")

#qogthingy<-data.frame(data_wide17[,1],ethnic10)
#colnames(qogthingy)[1]<-c("cntry")
#combined <- merge(x=qogthingy, y=combined, by="cntry")


qog_cs <-read_spss("C:/Users/Favone/Downloads/qog_std_cs_jan17.sav") #loading dataset
qog_cs$cntry<-countrycode(qog_cs$ccodecow, "cown","country.name")
qog_cs$legit3<-as.numeric(qog_cs$gov_ixlegitimacyindex)

legit_dat<-data.frame(qog_cs$cntry,qog_cs$legit3)
colnames(legit_dat)<-c("cntry","legit3")

aggrdelib <- merge(x=legit_dat, y=aggrdelib, by="cntry")

aggrdelib$legit3 <- range01(aggrdelib$legit3)
cor(na.omit(aggrdelib[,c(4,5,7,13)]))
cor(na.omit(aggrdelib[,c(4,5,2,6:13)]))

aggrdelib$legit <- 1-range01(aggrdelib$legit)
aggrdelib$legit2 <- 1-range01(aggrdelib$legit2)
aggrdelib$asia <- aggrdelib$e.asia + aggrdelib$s.e.asia + aggrdelib$s.asia + aggrdelib$pacific

SFI <-read_spss("C:/Users/Favone/Downloads/SFIv2016.sav") #loading dataset
SFI <- subset(SFI,SFI$year==2015)
SFI$cntry <-countrycode(SFI$country, "country.name","country.name") 
SFI$cntry[79] <- "North Korea"
SFI$cntry <-countrycode(SFI$cntry, "country.name","country.name") 
SFI$legitimacy <- SFI$legit
SFI$legit <- NULL


aggrdelib <- merge(x=SFI, y=aggrdelib, by="cntry")

hdi <- read_csv("hdi.csv")
hdi$cntry<-countrycode(hdi$Country, "country.name","country.name")
hdi$'1990'<- NULL ; hdi$'1991'<- NULL ; hdi$'1992'<- NULL ; hdi$'1993'<- NULL
hdi$'1994'<- NULL; hdi$'1995'<- NULL; hdi$'1996'<- NULL; hdi$'1997'<- NULL
hdi$'1998'<- NULL ; hdi$'1999'<- NULL ; hdi$'2011'<- NULL ; hdi$'2012'<- NULL
hdi$'2013'<- NULL; hdi$'2014'<- NULL; hdi$'2015'<- NULL;hdi$`HDI Rank (2015)`<- NULL
hdi$Country<- NULL
hdi10<-as.numeric(rowMeans(hdi[,1:11])) 
hdat<-data.frame(hdi$cntry,hdi10) 
names(hdat)<-c("cntry","hdi10")

aggrdelib <- merge(x=hdat, y=aggrdelib, by="cntry")
#table(aggrdelib$hdi10)
#gc()
#combined <- merge(x=hdat, y=combined, by="cntry")


#colnames(aggrdelib)[9]<-"e_p_polity"
#lop<-subset(vdems_start,vdems_start$year==2010)

cor(na.omit(aggrdelib2[,2:17]))
cor(na.omit(data.frame(aggrdelib$polity10,aggrdelib$delib10)))

#hist(vdems$v2dlconslt)
#hist(vdems$v2xcl_disc)
#table(vdems_start$e_boix_regime,vdems_start$year)
#table(vdems_sub$e_p_polity,vdems_sub$year)

##### merging time ####

combined <- merge(x=merged, y=aggrdelib, by="cntry")
table(combined$cntry)

combined <- as.data.frame(combined)
combined$gov_trust <- as.numeric(combined$gov_trust)
combined$age <- as.numeric(combined$age)
combined$income <- as.numeric(combined$income)
combined$educ <- as.numeric(combined$educ)
combined$sex <- as.factor(combined$sex)
combined$authoritarian <- as.numeric(combined$authoritarian)
combined$safety <- as.numeric(combined$safety)
# combined$demtoday <- as.numeric(combined$demtoday)
combined$latino <- factor(combined$latino)
combined$afro <- factor(combined$afro)
combined$americas <- factor(combined$americas)
combined$asia <- combined$e.asia + combined$s.e.asia + combined$s.asia + combined$pacific

cor(na.omit(data.frame(combined$gov_trust,combined$income,combined$educ, #Socioeconomic factors
                       combined$delib10, combined$polity10, combined$gdp10)))

combined$cntry<-as.factor(combined$cntry)


#combined$polity10 <- combined$polity10*20-10

#combined$regime <- combined$polity10
#combined$regime[combined$polity_autodummy==1] <- "auto"
#combined$regime[combined$polity_anodummy==1] <- "ano"
#combined$regime[combined$polity_demdummy==1] <- "demo"

hist(combined$gov_trust)
qplot(combined$gov_trust)

combined$cntry



#physi_s <- vdems_sub2 %>% dplyr::select(cntryears, franz)

#combined <- merge(combined, physi_s, by = "cntryears")

#physi2_s <- ddply(combined,~cntry,
#                summarise,politcat=mean(franz,na.rm=T))

#combined <- merge(combined, physi2_s, by = "cntry")


vdems_sub2$cntryears <- paste(vdems_sub2$cntry,vdems_sub2$year)

physi <- vdems_sub2 %>% dplyr::select(cntryears, perc)

combined <- merge(combined, physi, by = "cntryears")

physi2 <- ddply(combined,~cntry,
                     summarise,physviol=mean(perc,na.rm=T))

combined <- merge(combined, physi2, by = "cntry")





#physi_22$dis3 <-round(physi_22$dis2)

#unique(physi2_s$cntry)

#1-2.5
#3-5
#5.5 - 7
#physi2_s$politcat2<-8-((physi2_s$politcat) * (7/10))
#11 

#physi2_s$polity_demdummy <- physi2_s$politcat2
#physi2_s$polity_demdummy [physi2_s$politcat2 <= 2.5] <- 1
#physi2_s$polity_demdummy [physi2_s$politcat2 >  2.5] <- 0
#table(physi2_s$polity_demdummy)

#physi2_s$polity_anodummy <- physi2_s$politcat2
#physi2_s$polity_anodummy[physi2_s$politcat2 > 2.5 & physi2_s$politcat2 < 5.5] <- 1
#physi2_s$polity_anodummy[physi2_s$politcat2 <= 2.5 | physi2_s$politcat2 >= 5.5] <- 0
#table(physi2_s$polity_anodummy)

#physi2_s$polity_autodummy <- physi2_s$politcat2
#physi2_s$polity_autodummy[physi2_s$politcat2 >= 5.5] <- 1
#physi2_s$polity_autodummy[physi2_s$politcat2 < 5.5] <- 0
#table(physi2_s$polity_autodummy)

#physi2_s$regime <- physi2_s$politcat
#physi2_s$regime[physi2_s$polity_autodummy==1] <- "auto"
#physi2_s$regime[physi2_s$polity_anodummy==1] <- "ano"
#physi2_s$regime[physi2_s$polity_demdummy==1] <- "demo"

#unique(physi2_s$cntry)

table(physi2_s$regime)

qog2000 <- subset(qog,qog$year==2010 |
                       qog$year==2011 |  
                       qog$year==2012 | 
                       qog$year==2013 | 
                       qog$year==2014 | 
                       qog$year==2015)

table(qog2000$gd_ptsa,qog2000$year)

qog2000$perc2 <- qog2000$gd_ptsa

qog2000$cntry<-countrycode(qog2000$ccodecow, "cown","country.name")

qog2000$cntryears <- paste(qog2000$cntry,qog2000$year)

physi2000 <- qog2000 %>% dplyr::select(cntryears, perc2)

combined <- merge(combined, physi2000, by = "cntryears")

physi22000 <- ddply(combined,~cntry,
                summarise,terror=mean(perc2,na.rm=T))

combined <- merge(combined, physi22000, by = "cntry")

#table(qog2000$undp_hdi,qog2000$year)

#physiff <- qog2000 %>% dplyr::select(cntryears, undp_hdi)

#combined <- merge(combined, physiff, by = "cntryears")

#physiff2 <- ddply(combined,~cntry,
#                    summarise,hdi=mean(undp_hdi,na.rm=T))

#combined <- merge(combined, physiff2, by = "cntry")


gni <- read_csv("GNI.csv",  skip = 1)

gni$`2015`<-gsub("ttt","",gni$`2015`)
gni$`2015`<-gsub("ff","",gni$`2015`)
gni$`2015`<-gsub("sss","",gni$`2015`)
gni$`2015`<-gsub("uuu","",gni$`2015`)
gni$`2015`<-gsub("o","",gni$`2015`)
gni$`2015` <- as.numeric(gni$`2015`)

gni$`2014`<-gsub("ttt","",gni$`2014`)
gni$`2014`<-gsub("ff","",gni$`2014`)
gni$`2014`<-gsub("sss","",gni$`2014`)
gni$`2014`<-gsub("uuu","",gni$`2014`)
gni$`2014`<-gsub("o","",gni$`2014`)
gni$`2014` <- as.numeric(gni$`2014`)

gni<-gni[,c(2,23:28)]

gni<-gather(as.data.frame(gni),key = "Country")
names(gni) <- c("cntry","year","gni")

gni$cntry<-countrycode(gni$cntry,"country.name","country.name")
gni$cntryears <- paste(gni$cntry,gni$year)

physigni <- gni %>% dplyr::select(cntryears, gni)

combined <- merge(combined, physigni, by = "cntryears")

physigni2 <- ddply(combined,~cntry,
                    summarise,gni_c=mean(gni,na.rm=T))

combined <- merge(combined, physigni2, by = "cntry")

To Do - Regions

  mutate(regions = vdems_sub %>% 
           dcast(country_name ~ year, 
                 value.var=c("e_regionpol")) %>%
           select('2000') %>% 
           mutate(regions = fct_recode(as.factor(`2000`),
                                "E. Europe and C. Asia (post-Communist)" = "1",
                                "Latin America & Carribean" = "2",
                                "MENA" = "3",
                                "Sub-Saharan Africa" = "4",
                                "W. Europe and N. America" = "5",
                                "South & East Asia & Pacific" = "6",
                                "South & East Asia & Pacific" = "7",
                                "South & East Asia & Pacific" = "8",
                                "South & East Asia & Pacific" = "9",
                                "Latin America & Carribean" = "10")  %>%
                                as_factor))  %>%
           select(regions)


table(regions)

regions2 <- vdems_sub %>% 
  dcast(country_name ~ year, 
        value.var=c("e_regionpol")) %>%
  select('2000') %>% 
  transmute(fct_recode(as.factor(`2000`),
                       "E. Europe and C. Asia (post-Communist)" = "1",
                       "Latin America & Carribean" = "2",
                       "MENA" = "3",
                       "Sub-Saharan Africa" = "4",
                       "W. Europe and N. America" = "5",
                       "East Asia" = "6",
                       "South-East Asia" = "7",
                       "South Asia" = "8",
                       "Pacific" = "9",
                       "Latin America & Carribean" = "10"))

table(regions2)

regions3 <- dcast(vdems_sub, country_name ~ year, value.var=c("e_regionpol"))$"2000"

postcom   <- as.numeric(regions3 == 1)
latin     <- as.numeric(regions3 == 2 | regions3 == 10)
mena      <- as.numeric(regions3 == 3)
subsahara <- as.numeric(regions3 == 4)
west      <- as.numeric(regions3 == 5)
e.asia    <- as.numeric(regions3 == 6)
s.e.asia  <- as.numeric(regions3 == 7)
s.asia    <- as.numeric(regions3 == 8)
pacific   <- as.numeric(regions3 == 9)


ww <- vdems_sub %>% 
  dcast(country_name ~ year, 
                 value.var=c("e_regionpol")) %>%
  select('2000') %>% 
  mutate(regions = fct_recode(as.factor(`2000`),
                                "E. Europe and C. Asia (post-Communist)" = "1",
                                "Latin America & Carribean" = "2",
                                "MENA" = "3",
                                "Sub-Saharan Africa" = "4",
                                "W. Europe and N. America" = "5",
                                "South & East Asia & Pacific" = "6",
                                "South & East Asia & Pacific" = "7",
                                "South & East Asia & Pacific" = "8",
                                "South & East Asia & Pacific" = "9",
                                "Latin America & Carribean" = "10")) %>%
                                 as_factor() %>%
                                 select(regions)

ww

regions2 <- vdems_sub %>% 
  dcast(country_name ~ year, 
        value.var=c("e_regionpol")) %>%
  select('2000') %>% 
  transmute(fct_recode(as.factor(`2000`),
                       "E. Europe and C. Asia (post-Communist)" = "1",
                       "Latin America & Carribean" = "2",
                       "MENA" = "3",
                       "Sub-Saharan Africa" = "4",
                       "W. Europe and N. America" = "5",
                       "East Asia" = "6",
                       "South-East Asia" = "7",
                       "South Asia" = "8",
                       "Pacific" = "9",
                       "Latin America & Carribean" = "10"))

table(regions2)

regions3 <- dcast(vdems_sub, country_name ~ year, value.var=c("e_regionpol"))$"2000"

postcom   <- as.numeric(regions3 == 1)
latin     <- as.numeric(regions3 == 2 | regions3 == 10)
mena      <- as.numeric(regions3 == 3)
subsahara <- as.numeric(regions3 == 4)
west      <- as.numeric(regions3 == 5)
e.asia    <- as.numeric(regions3 == 6)
s.e.asia  <- as.numeric(regions3 == 7)
s.asia    <- as.numeric(regions3 == 8)
pacific   <- as.numeric(regions3 == 9)